darnit audit silently resolves upstream and never prints the audited repo

[justaugustus]

Description

darnit audit <path> auto-detects owner/repo preferring the upstream remote, and never prints which owner/repo it resolved. On a fork checkout (both origin and upstream present), the live checks target upstream — potentially a different repo than the one the user is working in — with no indication in the output.

Steps to Reproduce

1. Clone a fork so the checkout has both origin (your fork) and upstream remotes.
2. Run darnit audit <path> with authenticated gh.
3. Inspect which owner/repo the live gh api checks target.

Expected Behavior

The audit output states the resolved owner/repo and which remote it came from, so it's obvious which repository was audited.

Actual Behavior

The checks target upstream silently; the CLI logs only Auditing {path} with {N} controls. Details:

• packages/darnit/src/darnit/core/utils.py:186, :302 — prefer_upstream: bool = True.
• utils.py:231 — remotes = ["upstream", "origin"] if prefer_upstream else ["origin", "upstream"].
• cmd_audit calls detect_owner_repo(str(repo_path)) with no override and prints no repo identity.

The information already exists internally: detect_repo_from_git returns a source field (utils.py:263, 275), but the detect_owner_repo wrapper (:299) collapses it to (owner, repo) and the CLI discards it.

Environment

• darnit version: main @ 86f9cfa
• Python version: 3.12 (uv-managed)
• OS: any

Additional Context

Suggested fix (visibility only): print the resolved owner/repo + source remote at the top of audit output. A new --owner/--repo flag is likely unnecessary — the origin-preference path (prefer_upstream=False) already resolves origin correctly.